Materials and products must be inspected to assure that the quality characteristics conform to requirements. Inspection may occur as the product is being produced, at final Inspection on completed product at the producer, or at receiving Inspection at the consumer. There are three ways that a lot of N quantity may be Inspected: complete Inspection of the lot, no Inspection of the lot, or a partial inspection of the lot. Complete inspection of the lot, referred to as 100% inspection, can be extremely time consuming, and as such expensive. 0% inspection should be reserved for those situations where even a single defect is associated with unacceptable risk, as in medical or aerospace applications. 100% inspection may also be necessary if there is reason to believe that the lot is of particularly low quality, or if no information is available to estimate the lot quality. No inspection of the lot, or 0% inspection, is the ideal inspection level from a financial view, as there is no cost added. However, 0% inspection Is risky, as even one bad lot of material can have a significant monetary impact, easily erasing any savings realized from the lack of Inspection.
Regardless, certain situations do lend themselves to Inspection. The material may be so Inexpensive, Like a screw or nut, that there Is no Justification for Inspection. Or there may be sufficient statistical and/or historical evidence that the lot will meet the required quality level that inspection is unwarranted. A partial inspection of the lot, called sample inspection, provides an alternative to the extremes of 100% or 0% inspection, and is the most common method of lot inspection. Sample inspection Lana typically use statistically derived tables from a know standard, such as the ubiquitous IEEE military standard.
These sampling plans allow for the selection of an Acceptable Quality Level (SQL) with a corresponding sample size (n) based on lot size (N). The inspector then uses the specified acceptance number (c) and rejection number (d or r) to decide if the lot should be accepted or rejected, called sentencing the lot. The lot Is accepted and considered of adequate quality when the number of Identified defects In the sample is less than or equal to the acceptance number; otherwise the lot Is rejected. A refinement to single sampling plans, where a single sample is used for lot sentencing, is a double sampling plan.
Instead of a single sample (ml), a second sample size (no) is also defined by the selected plan, as well as a second set of acceptance and rejection numbers (ca and do). If the number of defects identified in the first sample is less than CLC, the lot is accepted; if the defects are greater than ca, the lot is rejected. If the number of defects is greater than CLC, but less than or equal to ca, a second sample is drawn. If the sum of the defects identified in both samples is less than or equal to ca, the lot is accepted, if the sum is greater than ca, the lot Is rejected.
Double sampling plans can be psychological appealing, with a perceived “second chance” for accepting a lot. The reality Is that no such advantage exists, as both single and double sampling plans are designed to provide similar probable’s for accepting or rejecting lots of Identical quality. The actual advantage of double sampling plans over single sampling plans Is found In the sampling plan will always have a lower ASSN then a single sampling plan; the inspection took less time and so cost less.